xzxzxzxz/Awesome-What-Bimanual-Can-Do

30

37 commits

updated Dec 16, 2025

See the code

README

Awesome-What-Bimanual-Can-Do

WBCD Logo

arXiv Home Page LIVE X

Awesome visitors Maintenance PR's Welcome

Media Coverage (CN): 甲子光年 | 36氪 | 机器之心 | 量子位

Hardware Sponser's Highlight: Galaxea 星海图(Track#1) | AgileX 松灵机器人(Track#2) | ARX 方舟无限(Track#3)

😎 A curated list of research, news, datasets, benchmarks, hardware systems, challenges/competitions and real-world applications in Bimanual Manipulation and Learning from Demonstrations — affiliated with the ICRA 2025 What Bimanual Teleoperation and Learning from Demonstration Can Do (WBCD) Today Competition, founded by Dr. Zhuo Xu. Stay tuned as this list evolves alongside the upcoming technical survey, which will include taxonomy, insights, and contributions within recent 5 years and our lessons learned from WBCD competition.

📣 Disclaimer: The paper release timestamps and acceptance status are primarily based on the comment section on arXiv. We will continue to verify and update this information. If you are the author of a paper, feel free to open an issue to help us keep things up to date!

📣 Call of PRs! Think we missed your awesome paper or work? You're very welcome to open a pull request with the title and link — we'll review and update after checking the scope!

Table of Contents


😶‍🌫️ 0. Our Vision

In recent years, the field of bimanual teleoperation and learning from demonstration (LfD) has witnessed remarkable progress, driven by innovations in human sensing technologies and diverse robotic embodiments. Systems like ALOHA, Mobile ALOHA, DexCap, Open-TeleVision, HATO, GELLO, AirExo, and UMI have demonstrated the potential of combining intuitive control interfaces with data-driven learning techniques to enable dexterous, human-like robot manipulation.

However, much of the existing research has been conducted in isolated settings, with researchers designing their own tasks and evaluating their own systems. Key industrial concerns such as motion efficiency, system robustness, cost-effectiveness, and the ease of learning deployable policies—have often been overlooked in academic benchmarks.

To bridge this gap between research and real-world impact, we launched the ICRA 2025 WBCD Competition. The competition aimed to create a shared platform where the community could evaluate technologies on a common set of challenging, practical, and industry-relevant tasks, using diverse human-sensing and teleoperation methods (e.g., VR, puppeteering, exoskeletons, hand-held grippers, gloves and more).

This awesome list is a continuation of that effort. Our vision is to curate, consolidate, and continuously update the most impactful work in the field from the past five years—covering algorithms, datasets, simulators, and hardware systems—and to foster collaboration across academia and industry toward building practical, generalizable bimanual manipulation.


🎟️ 1. Events



📑 3. WBCD - Technical Paper List

Includes accepted and arXiv technical papers on bimanual manipulation.
🚧 TODO: Currently listed in reverse chronological order by year; further categorization by algorithmic approach and end-effector type is in progress.

2025
2024
2023
2022
2021

📊 4. WBCD - Dataset & Benchmark

Includes existing datasets and benchmarks that support bimanual manipulation.
🚧 TODO: Currently listed in reverse chronological order by year; further categorization is in progress.

2025

🧪 5. WBCD - Simulator

2025
2024

🤖 6. WBCD - Hardware

Includes hardware systems designed for bimanual manipulation tasks.
🚧 TODO: Currently listed in reverse chronological order by year; further classification into lab prototypes vs. commercial/industrial products is in progress.

2025

🕹️ 7. WBCD - Data Collection Solution

Includes various data collection setups and methods for bimanual manipulation tasks.
🚧 TODO: TBD

2025

xzxzxzxz/Awesome-What-Bimanual-Can-Do

30

37 commits

updated Dec 16, 2025

See the code

README

Awesome-What-Bimanual-Can-Do

WBCD Logo

arXiv Home Page LIVE X

Awesome visitors Maintenance PR's Welcome

Media Coverage (CN): 甲子光年 | 36氪 | 机器之心 | 量子位

Hardware Sponser's Highlight: Galaxea 星海图(Track#1) | AgileX 松灵机器人(Track#2) | ARX 方舟无限(Track#3)

😎 A curated list of research, news, datasets, benchmarks, hardware systems, challenges/competitions and real-world applications in Bimanual Manipulation and Learning from Demonstrations — affiliated with the ICRA 2025 What Bimanual Teleoperation and Learning from Demonstration Can Do (WBCD) Today Competition, founded by Dr. Zhuo Xu. Stay tuned as this list evolves alongside the upcoming technical survey, which will include taxonomy, insights, and contributions within recent 5 years and our lessons learned from WBCD competition.

📣 Disclaimer: The paper release timestamps and acceptance status are primarily based on the comment section on arXiv. We will continue to verify and update this information. If you are the author of a paper, feel free to open an issue to help us keep things up to date!

📣 Call of PRs! Think we missed your awesome paper or work? You're very welcome to open a pull request with the title and link — we'll review and update after checking the scope!

Table of Contents


😶‍🌫️ 0. Our Vision

In recent years, the field of bimanual teleoperation and learning from demonstration (LfD) has witnessed remarkable progress, driven by innovations in human sensing technologies and diverse robotic embodiments. Systems like ALOHA, Mobile ALOHA, DexCap, Open-TeleVision, HATO, GELLO, AirExo, and UMI have demonstrated the potential of combining intuitive control interfaces with data-driven learning techniques to enable dexterous, human-like robot manipulation.

However, much of the existing research has been conducted in isolated settings, with researchers designing their own tasks and evaluating their own systems. Key industrial concerns such as motion efficiency, system robustness, cost-effectiveness, and the ease of learning deployable policies—have often been overlooked in academic benchmarks.

To bridge this gap between research and real-world impact, we launched the ICRA 2025 WBCD Competition. The competition aimed to create a shared platform where the community could evaluate technologies on a common set of challenging, practical, and industry-relevant tasks, using diverse human-sensing and teleoperation methods (e.g., VR, puppeteering, exoskeletons, hand-held grippers, gloves and more).

This awesome list is a continuation of that effort. Our vision is to curate, consolidate, and continuously update the most impactful work in the field from the past five years—covering algorithms, datasets, simulators, and hardware systems—and to foster collaboration across academia and industry toward building practical, generalizable bimanual manipulation.


🎟️ 1. Events



📑 3. WBCD - Technical Paper List

Includes accepted and arXiv technical papers on bimanual manipulation.
🚧 TODO: Currently listed in reverse chronological order by year; further categorization by algorithmic approach and end-effector type is in progress.

2025
2024
2023
2022
2021

📊 4. WBCD - Dataset & Benchmark

Includes existing datasets and benchmarks that support bimanual manipulation.
🚧 TODO: Currently listed in reverse chronological order by year; further categorization is in progress.

2025

🧪 5. WBCD - Simulator

2025
2024

🤖 6. WBCD - Hardware

Includes hardware systems designed for bimanual manipulation tasks.
🚧 TODO: Currently listed in reverse chronological order by year; further classification into lab prototypes vs. commercial/industrial products is in progress.

2025

🕹️ 7. WBCD - Data Collection Solution

Includes various data collection setups and methods for bimanual manipulation tasks.
🚧 TODO: TBD

2025